Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
3D voxel data analysis of silver nanowire meshes in stretchable electronics
3D X-ray reconstructed tomographies of silver nanowires embedded in PolyDiMethylSiloxane was collected on the TOMCAT line at the Swiss Light Source at 65^3 nm^3 voxel size. The position and orientation of wires need to be extracted from data through various image analysis approaches.
Stretchable electronics is a booming field due to the variety of biomedical, neuroscience and consumer electronics applications. Implantable strain sensors for bladder monitoring in paraplegic patients, brain-shape-adapting microelectrode arrays for brain signal mapping, and flexible screens, phones are just a few of thousands of applications. Most of these technologies have at their core the same principle : conductive nanowires moving on top of each other embedded in a deformable, usually plastic matrix.
From a modeling perspective, to understand the electrical behavior of these systems, we lack the complete orientation and stacking statistics of these wires. Furthermore, fatigue and hysteresis and short-scale viscoelastic relaxation are common yet poorly understood effects in this technology.
X-ray reconstructed tomography allows for fine inspection of nanowires inside these samples. At the Swiss Light Source, the TOMCAT Line provides the worlds largest field of view (70^3 um^3) and best resolution at a voxel size of 65^3 nm^3 ( therefore ~180nm resolution ) at the used energy (13keV). We have collected data on various samples made of Silver but also Gold nanowires, ranging from 150 to 300 nm in diameter an tens of microns in length. Samples were stretched in-situ, and deformation of the same group of wires could be tracked. Lastly, fast pull tests were performed to investigate the viscoelastic relaxation of these samples.
The human eye can make out the wires in the collected data, but it is challenging to segment the 3D space to extract individual wires. This is necessary to assess the number of contacts per wire, their spatial orientation distribution and the stacking of the wires.
Several approaches are employed. Common segmentation processes fail. Work will be mostly done in Python, with bindings to self-written (or available) C-Code, and also for perfomance (GPU) CUDA Code. Prior knowledge is better, but not necessary.
Stretchable electronics is a booming field due to the variety of biomedical, neuroscience and consumer electronics applications. Implantable strain sensors for bladder monitoring in paraplegic patients, brain-shape-adapting microelectrode arrays for brain signal mapping, and flexible screens, phones are just a few of thousands of applications. Most of these technologies have at their core the same principle : conductive nanowires moving on top of each other embedded in a deformable, usually plastic matrix.
From a modeling perspective, to understand the electrical behavior of these systems, we lack the complete orientation and stacking statistics of these wires. Furthermore, fatigue and hysteresis and short-scale viscoelastic relaxation are common yet poorly understood effects in this technology.
X-ray reconstructed tomography allows for fine inspection of nanowires inside these samples. At the Swiss Light Source, the TOMCAT Line provides the worlds largest field of view (70^3 um^3) and best resolution at a voxel size of 65^3 nm^3 ( therefore ~180nm resolution ) at the used energy (13keV). We have collected data on various samples made of Silver but also Gold nanowires, ranging from 150 to 300 nm in diameter an tens of microns in length. Samples were stretched in-situ, and deformation of the same group of wires could be tracked. Lastly, fast pull tests were performed to investigate the viscoelastic relaxation of these samples.
The human eye can make out the wires in the collected data, but it is challenging to segment the 3D space to extract individual wires. This is necessary to assess the number of contacts per wire, their spatial orientation distribution and the stacking of the wires.
Several approaches are employed. Common segmentation processes fail. Work will be mostly done in Python, with bindings to self-written (or available) C-Code, and also for perfomance (GPU) CUDA Code. Prior knowledge is better, but not necessary.
Semester project (not optimal): reconstruction and preliminary tests of several algorithms to improve image contrast
Master project: Successful segmentation of the data, Physics simulations using Bullet Physics on the GPU to simulate wire stacking.
Semester project (not optimal): reconstruction and preliminary tests of several algorithms to improve image contrast
Master project: Successful segmentation of the data, Physics simulations using Bullet Physics on the GPU to simulate wire stacking.
Csaba Forro
forro@biomed.ee.ethz.ch ETZ F75, Gloriastrasse 35 8092 Zurich
The lab is cool. If you're interested, send me an e-mail, drop by for a coffee. We can negotiate.